جستجوی مقالات مرتبط با کلیدواژه « Wavelet Transform » در نشریات گروه « آب و خاک »

Predicting river flow is one of the most crucial aspects in water resources management. Improving forecasting methods can lead to a reduction in damages caused by hydrological phenomena. Studies indicate that artificial neural network models provide better predictions for river flow compared to physical and conceptual models. However, since these models may not offer reliable performance in estimating unstable data, using preprocessing techniques is necessary to enhance the accuracy and performance of artificial neural networks in estimating hydrological time series with nonlinear relationships. One of these methods is wavelet transformation, which utilizes signal processing techniques.

In this study, to evaluate the efficiency of discrete and continuous wavelet types in the Wavelet-Artificial Neural Network (WANN) hybrid model for monthly flow prediction, a case study was conducted on the Kardeh Dam watershed in the northeast of Iran, serving as a water source for part of Mashhad city and irrigation downstream agricultural lands. Monthly streamflow estimates for the upstream sub-basin of the Kardeh Dam were obtained from the meteorological and hydrometric stations' monthly statistics over a 30-year period (1991-2020). The WANN model is a hybrid time series model where the output of the wavelet transform serves as a data preprocessing method entering an artificial neural network as the predictive model. The combination of wavelet analysis and artificial neural network implies using wavelet capabilities for feature extraction, followed by the neural network to learn patterns and predict data, potentially enhancing the models' performance by leveraging both methods. The 4-fold cross-validation method was employed for the artificial neural network model validation, where the model underwent validation and accuracy assessment four times, each time using 75% of the data for training and the remaining 25% for model validation. The final results were presented by averaging the validation and accuracy results obtained from each of the four model runs. To evaluate and compare the performance of the models used in this study, three evaluation indices, Nash-Sutcliffe Efficiency (NSE), Root Mean Square Error (RMSE), and Pearson correlation coefficient (R), were employed.

The analysis of meteorological and hydrometric data in this study revealed that monthly streamflow in two time steps, T-1 and T-2, were the most effective predictive variables. Each of the two runoff variables of the previous month (Qt-1) and the previous two months (Qt-2) were analyzed by each of the Haar and Fejer-Korovkin2 discrete wavelet transforms and the two continuous Symlet3 and Daubechies2 wavelets at three levels. The results of each level of decomposition was given as input to the ANN model. The presented results at each decomposition level indicated that hybrid models could accurately predict lower flows compared to the single ANN model, and the estimation of maximum values also significantly improved in the hybrid models. Among the wavelets used, Haar wavelets exhibited the weakest performance, and the less commonly employed Kf2 wavelet showed a moderate performance. Since the Haar and Fk2 wavelets, with their discrete structure, did not perform well in decomposing continuous monthly streamflow data, continuous wavelet models outperformed discrete wavelet models. The hybrid models, combining wavelet analysis and artificial neural networks, demonstrated up to an 11% improvement over the performance of the single neural network model.

Streamflow is a crucial element in the hydrological cycle, and predicting it is vital for purposes such as flood prediction and providing water for consumption. The objective of this research was to evaluate the performance of different types of discrete and continuous wavelet models at various decomposition levels in enhancing the efficiency of artificial neural network (ANN) models for streamflow prediction. Since climate and watershed characteristics can influence the nature of data fluctuations and, consequently, the results of the wavelet model decomposition, choosing an appropriate wavelet model is essential for obtaining the best results. Considering the existing variations in the results of different studies regarding the selection of the best wavelet type, it is suggested to use both continuous and discrete wavelet types in modeling to achieve the best predictions and select the optimal results. Given that a lower number of input variables in neural network models lead to higher accuracy in modeling results, it is recommended to perform decomposition at a two-level depth to reduce input components to the neural network model, thereby reducing the model execution time.

Using the values analyzed by the wavelet function can increase the accuracy of the simulations. Considering the climatic changes and the increase of extreme values in recent years, in this study, we made an effort that the effect of signal processing under the name of wavelet transformation in improving the performance of random forest model in simulating monthly river flow in Siminehrood and Mahabadchai sub-basins in the south of Lake Urmia has been discussed and investigated in the period of 1971-2019.

In this study, the accuracy of the random forest model has been investigated in two steps of training and testing. At first, the random forest model was evaluated in two phases of training and testing in rainfall-runoff simulation in the south of Lake Urmia basin. Nash-Sutcliffe statistics and root mean square error were used to evaluate the performance and error rate of the studied models, respectively. In the next step, after investigating the performance of the random forest model, the time series of rainfall and river flow in the studied basins were analyzed using the wavelet function. In this regard, two analysis levels (level 1 and 2) and two Haar and Daubechies wavelet functions were used. Finally, using the random forest model, rainfall-runoff simulation based on the wavelet theory was done under the name of W-RF model.

Results and Dissection: 

At First, the random forest model was investigated in two phases of training and testing, and the simulation results of the river flow values showed that the simulated values were within the 95% confidence interval, and the error rate of the river flow simulation using the RMSE statistic is 3.22 and 8.91 cubic meters per second in the test phase for Mahabadchai and Siminehrood sub-basins, respectively. In order to investigate the effect of time series analysis on the performance of the RF model, wavelet theory and Haar and Daubechies 4 wavelet functions were used in decomposition levels 1 and 2. By estimating the accuracy and performance of the hybrid W-RF model in 4 input patterns, the best pattern was selected based on the RMSE and NSE model evaluation criteria. The research results showed that for the Haar wavelet function in level 1 decomposition has better performance and error rate than level 2 type in both sub-basins. In this study, the Daubechies wavelet at level 1 in the test phase has provided the best performance and the lowest error rate in the simulation of the river flow values in the studied sub-basins and has been able to reduce the error rate in the two sub-basins of Mahabadchai and Siminehrood respectively by about 89 and 80 percent compared to the random forest model.

Finally, by comparing the RF and W-RF models, the simulation results of river flow in the two studied sub-basins showed that the integrated W-RF model was able to reduce the error rate in the two sub-basins of Mahabadchai and Siminehrood to reduce by 89 and 80% respectively. Considering the increase in simulation complexity with the involvement of wavelet theory, the error recovery rate and model performance are acceptable. The integrated W-RF model in this study provides reliable results for the simulation of river flow data in order to support decision-making and risk analysis in the exploitation of downstream reservoirs and the management of water resources in sub-basins. The obtained results can be used in the design of water resources systems.

Iran is faced with a dry and semi-dry climate with destructive floods, droughts, and water shortages. Droughts and floods can affect the environment, economic, and social activities. Therefore, examining and predicting river discharge and planning management to control it, especially for future water consumption, is very valuable. In this study, changes in river discharge were modeled using statistical data from 2001 to 2020.Statistical data from synoptic and hydrometric stations in a semi-arid region in Urmia city of West Azerbaijan province and a humid region in Amol city of Mazandaran province were used. Out of twelve time-series models defined for the Long Short-Term Memory (LSTM) network, the best model was identified. Then, LSTM modeling was performed based on pre-processing methods of Discrete Wavelet Transform (DWT) and Complementary Ensemble Empirical Mode Decomposition (CEEMD). The results showed that the selected model has high ability and efficiency in estimating the amount of river discharge. On the other hand, pre-processing methods improved the results such that the DC evaluation criterion in the wavelet transform increased from 0.93 to 0.95 in the Nazloo River and from 0.83 to 0.90 in the Chalous River. The best evaluation results for test data using wavelet transform for the Nazloo River in the semi-arid climate with evaluation criteria of R=0.977, DC=0.954, and RMSE=0.018 were obtained. Furthermore, the results of the sensitivity analysis indicated that the discharge parameter of one day before is the most effective in daily discharge estimation.

In this study, the effect of empirical mode decomposition (EMD) and maximal overlap discrete wavelet transform (MODWT) approaches in the processing of average temperature, wind speed and precipitation annual, monthly and single month time scales were evaluated. The selected study stations were Ramsar, Kermanshah, Hamadan, Semnan, Mashhad, Shiraz, Ardabil,Urmia, Zanjan, Yazd, Tabriz, Maragheh and Mianeh. In the time series decomposition using EMD, the trend of combined sub-series and in MODWT, the trend of sub-series was consistent with the Mann-Kendall test values of observed data time series, although, it can also be affected with the value and sign of of Mann-Kendall statistic. The significance level of The Mann-Whitney test in comparison of main annual observed time series with subseries of both approaches was more than 0.05, which indicates the acceptable performance of EMD and MODWT. In general, the comparisons of the sub-series of two approach with the main time series in each station, showed that the corresponding significance level for MODWT approach is higher than EMD. Also, the average of significance level for temperature series greater than those of precipitation and wind speed series. The LSD test at the 95% confidence level indicates good agreement of the observed data time series and combined sub-series.

Water pollution is a major global problem that requires constant evaluation and revision of water resources policy at all levels. Dissolved oxygen (DO) is one of the most important indicators of water quality. In the present study, the water quality parameter of dissolved oxygen using intelligent Long Short-Term Memory (LSTM) method based on discrete wavelet transform (DWT) and Complementary Ensemble Empirical Mode Decomposition (CEEMD) pre-processor methods in both temporal and spatial modes. It was investigated in five consecutive stations on the Savannah River. The results of analysis of models showed the ability and high efficiency of the method used in estimating the amount of dissolved oxygen in water. On the other hand, pre-processor methods improved the results. It was also observed in the investigations that the results of analysis based on wavelet transformation in spatial modeling reduced the RMSE error by two percent and also the empirical mode decomposition in temporal modeling by 15 percent. The best evaluation for test data was obtained using the empirical mode decomposition in temporal modeling corresponding to the previous day with values ​​of DC=0.977, R=0.988 and RMSE=0.017. Also, in the spatial modeling to estimate dissolved oxygen in the third station, it was found that the results obtained from the inputs of the dissolved oxygen parameter one day before the second station and two days before the first station have the best results.

Surface and underground waters are one of the world's most important problems and environmental concerns. In the last few decades, due to the rapid growth of the population, the water needs have increased, followed by the input load to the water. In order to classify the quality of underground water and water level according to the type of consumption, there are many methods, one of the most used methods is the use of quality indicators. Considering the facilities available in water quality monitoring stations and the need to save time and money, using alternative methods of modern data mining methods can be good for predicting and classifying water quality. The process of water extraction for domestic use, agricultural production, mineral industrial production, electricity production, and ester methods can lead to the deterioration of water quality and quantity, which affects the aquatic ecosystem, that is, the set of organisms that live and interact. Therefore, it is very important to evaluate the quality of surface water in water-environmental management and in monitoring the concentration of pollutants in rivers. The aim of the current research was to estimate the numerical values of the drinking water quality index (WQI) using the tree method and investigate the effect of wavelet transformation, the Bagging method, and principal component analysis.
In this research, to calculate the WQI index from the quality parameters of the Bagh Kalaye hydrometric station including total hardness (TH), alkalinity (pH), electrical conductivity (EC), total dissolved solids (TDS), calcium (Ca), sodium (Na), Magnesium (Mg), potassium (K), chlorine (Cl), carbonate (CO3), bicarbonate (HCO3) and sulfate (SO4) were used in the statistical period of 23 years (1998-2020). Quantitative values calculated with the WQI index were considered as target outputs. By using the relief and correlation method, the types of input combinations were determined. The random tree method was used to estimate the numerical values of the WQI index. Then, the capability of the combined approach of wavelet, principal component analysis, and Bagging method with random tree base algorithm was evaluated. To compare the values obtained from the data mining methods with the values calculated from the WQI index, the evaluation criteria of correlation coefficient (R), root mean square error (RMSE), mean absolute error (MAE), and modified Wilmot coefficient (Dr) were used.
The use of the wavelet transform method and the Bagging method has improved the modeling results. Considering that the Bagging classification method with the random tree base algorithm is a combination of the results of several random trees, so using this method has increased the accuracy of the RT model. So, in general, it was concluded that the use of wavelet transformation and classification methods increases accuracy and reduces errors. The best scenario with the highest accuracy and the lowest error was related to scenario 10 of the W-B-RT model with Total Hardness, Electrical Conductivity, Total Dissolved Solid, Sulphate, Calcium, Bicarbonate, Magnesium, Chlorine, Sodium, and potassium parameters. The results showed that the effect impact of pH in estimating the numerical value of the WQI index is considered lower than other parameters. When the principal component analysis method was used, by reducing the value of the eigenvalue from F1 to F12, the value of the factor also decreased; As a result,so F1, F2, and F3 factors were selected as the basic components. Considering 3 main factors, modeling was done employed and R=0.98, RMSE=2.17, MAE=1.52, and Dr=0.97 were obtained. In general, the results showed that the PCA method, despite reducing the dimension of the input vectors and simplifying it, can improve the accuracy and speed of the model and is introduced as the best method for estimating the numerical value of the WQI index.
The results obtained from the present research showed that the use of wavelet transform, Bagging and PCA methods had a positive effect on improving the results and increasing higherthe accuracy. In estimating the numerical values of WQI index, PCA-B-RT method considering 3 main factors, with correlation coefficient equal to 0.98, root mean square error equal to 2.17, average absolute value error equal to 1.52 and tThe modified Wilmot coefficient equal to 0.97 had the highest accuracy. Considering that all the methods used in the estimation of quantitative values had acceptable accuracy, therefore, in case of lack of data and lack of access to all chemical parameters, it is possible to obtain appropriate and acceptable results by using a limited number of parameters and data mining methods achieved.

Estimation and simulation of precipitation is considered as one of the most issues in the field of hydrology. In this study, for the first time, the long-term rainfall in Babolsar city during a 68 years period from 1951 to 2019 was predicted by using an optimized hybrid artificial intelligence (AI) technique. To end this, the gene expression programming (GEP) model was combined with the wavelet transform. To training the AI models, 70% of the observed values were utilized and 30% of these values were used to testing those models. Additionally, the autocorrelation function (ACF) was applied to identify the most influential lags and then six GEP models were developed by means of these detected lags. The number of optimized genes was selected to be four. In addition, the Multiplication function was introduced as the best linking function of the GEP model. The superior GEP model was introduced through a sensitivity analysis that the correlation coefficient (R) and scatter index (SI) for this model were calculated to be 0.571 and 0.792, respectively. The (t-1), (t-2), (t-12), and (t-36) time series lags were introduced as the most effective input lags. The coif was detected as the best mother wavelet to simulate the target function. The hybrid WGEP model simulated the values of rainfall with acceptable accuracy. In the other words, the wavelet transform enhanced the performance of the GEP model significantly. For instance, the value of variance accounted for (VAF) for the GEP and WGEP models were respectively computed to be 31.710 and 82.064

Unpredictable settlement of earth dams has led researchers to develop new methods such as artificial neural networks, wavelet theory, fuzzy logic, and a combination of them. These methods do not require time-consuming analyses for estimation. In this research, the amount of settlement in rockfill dams with a central core has been estimated using artificial intelligence methods. The data of 35 rockfill dams with a central core were used to train and validate the models. The artificial neural network, wavelet transform model, and fuzzy-neural adaptive inference system are the proposed models which were used in the present study. According to the results, the best model for an artificial neural network had two hidden layers, the first layer of 18 neurons and the second layer of 7 neurons, with the Tansig-Tansig activation function, with a coefficient of determination R2=0.4969. The best model for the fuzzy-neural inference system had the ring function (Dsigmoid) as a membership function, with three membership functions and 142 repetitions with a coefficient of determination R2=0.2860. Also, combining wavelet-neural network conversion with the coif2 wavelet function due to the more adaptation this function has to the input variables, the better the performance, and this function, with a coefficient of determination R2=0.9447, had the highest accuracy compared to other models.

In this study, the long-term rainfall in Babolsar city was simulated using an optimized hybrid artificial intelligence (AI) model over a 68 years period during 1951 to 2019. The ANFIS network and the wavelet transform combined to develop the hybrid model. Firstly, the effective lags of time series data were identified by using the autocorrelation function (ACF). Six ANFIS models were defined through these lags, Subsequently. The best ANFIS model was detected by conducting a sensitivity analysis. For testing the best ANFIS model, the value of correlation coefficient (R), variance accounted for (VAF), and scatter index (SI) were respectively computed to be 0.612, 37.029, and 0.761. In Addition, results of the models analysis showed that the (t-1), (t-2), (t-12), and (t-36) were identified as the most significant lags. Finally, the superior hybrid model was examined in three decomposition level (DL), revealing that the best results were obtained from the second decomposition level (DL2). In testing mode of the model, the R, VAF, and SI were calculated to be 0.972, 94.455, and 0.266, respectively. Therefore, the simulation results showed that the wavelet transform enhanced the performance of the ANFIS network significantly.

The aim of this study was to estimate the groundwater level of Aleshtar plain at times t + 1, t + 3 and t + 6 using the parameters of temperature, precipitation and groundwater level at times t, t-1 and t-2 using models: neural network (ANN) , a neural fuzzy inference system (ANFIS), a neural-wavelet (WNN), and an integrated neural-fuzzy-wavelet (WNF) network. Two indices R2 and RMSE were used to evaluate the models. The results of predicting different models showed that ANFIS, WNN and WNF have higher accuracy in predicting groundwater level than ANN model. Also, the comparison of the results obtained from wavelet-based models and other models showed that these models (WNN and WMF) have higher accuracy than other models due to pre-processing and data analysis. The use of WNF model compared to ANN has increased the R2 index from 0.94 to 0.98 (in one-month forecast), 0.84 to 0.93 (in three-month forecast) 0.76 to 0.85 ( In the six-month forecast). Also, WNF compared to ANN model, has decreased RMSE index from 0.56 to 0.32 (in one-month forecast), 0.96 to 0.66 (in three-month forecast) and 1.18 to 0.97 (in the six-month forecast). The results of groundwater depth prediction with four models showed that these models have more accurate in predicting shorter time steps. Also, using of models in predictions with a time delay of more than three months, not only does not have much effect on the accuracy of the model, but in models with wavelet basis reduces the accuracy.

In this study, the long-term rainfall in Babolsar city was simulated by means of an optimized AI model. To do this, the extreme learning machine (ELM) and the wavelet transform (WT) were combined. It should be stated that the monthly rainfall values from 1951 to 2019 were applied, meaning that 70% of the observed values were employed to training the AI models and 30% of rest were utilized to testing these models. Firstly, the activation functions of the ELM models were evaluated; as a result, the sigmoid was chosen as the best activation function. Moreover, the lags of time series were introduced using the autocorrelation function (ACF) that four ELM models were defined through those identified lags. By performing a sensitivity analysis, the superior ELM model was introduced. The values of correlation coefficient (R), variance accounted for (VAF), and scatter index (SI) for the ELM model were respectively computed to be 0.524, 27.064, and 0.819. Furthermore, different mother wavelets were examined and the “dmey” was opted as the best mother wavelet. The wavelet transform enhanced the accuracy of the simulations significantly. For instance, the VAF index for the hybrid WELM model equaled to 86.461. It is noteworthy that the hybrid model was evaluated for different decomposition levels (DL) and then the best one was detected. Also, the (t-1) and (t-12) lags were identified as the most effective input lags.

This study analyses the trends in runoff and rainfall data at monthly timescales observed at three rainfall and one runoff stations of Ardabil plain in 1350-1394 period. The purpose of this study is to analyse the trends in runoff and rainfall time series using MK1, MK2 and MK3 tests. Also, in order to determine the dominant periodic components applied hybrid Wavelet Transform and MK1 test. The results of MK1, MK2 and MK3 tests showed that decreasing trend is prevalent in rainfall (P1) and runoff (R1) of Samian station and positive trend in rainfall of Gilandeh (P2) and Kozatopraghi (P3) stations. Also, rainfall time series have not shown non-significant trend. But runoff time series were shown significant trend. The results of MK after removing the effect of all significant serial correlation (MK3) showed the value of trends decreased in comparison to MK1 parameters. Generally, short-term periods were important in the involved trend at original time series. Thus, the rainfall data showed short-term periods of 2 months (A+D1) in monthly scales, respectively. In the runoff time series, wavelet-based detail at level 3 plus the approximation (A+D3) time series was conceded as the dominant periodic component.

Baseflow separation is one of the important topics in hydrology and water resources management. Different methods have been used to baseflow separation in the river and karstic springs. In this study, the wavelet transform was used as a new tool to baseflow separation in Gamasiyab spring in Nahavand County and Kahman spring in Aleshter County. In this study the data of the 49 years period for the Gamasiyab spring and data of the 27 years period for the Kahman spring were used. For evaluation of this tool, the Hydrochemical method was used by sampling quality data in water year of 2017-2018 for the two springs. The results showed that, in Gamasiyab spring, wavelet function Db4 at level 6 has the best performance. This best performance has the adjusted R2 of 0.96, and RMSE of 0.651 (m3/s). The best performance in Kahman spring related to wavelet function Db2 at level 5, that has adjusted R2 of 0.99, and RMSE of 0.187 (m3/s). Generally, the Wavelet Transform has high accuracy in baseflow separation.

Rainfall may be considered as the most important source of drinking water and watering land in different areas all over the world. Therefore, simulation and estimation of the hydrological phenomenon is of paramount importance. In this study, for the first time, the long-term rainfall in Rasht city was simulated using an optimum hybrid artificial intelligence (AI) model over a 62 year period from 1956 to 2017. The gene expression programming (GEP) and wavelet transform (WT) were combined to develop the hybrid AI model (WGEP). Firstly, the most effective lags of time series data were identified by means of the autocorrelation function (ACF); then eight various GEP and WGEP models were defined. Next, the GEP models were analyzed and the superior GEP model as well as the most influenced lags was detected. For instance, the variance accounting for (VAF), correlation coefficient (R) and scatter index (SI) for the superior GEP model was calculated to be 0.765, 0.508 and 0.709, respectively. Additionally, lags (t-1), (t-2), (t-3) and (t-12) were the most influenced. Then, the different mother wavelets were examined, indicating that the demy mother wavelet was the most optimal one. Moreover, analyzing the numerical simulations showed that the mother wavelet enhanced the performance of the GEP model significantly. For example, the VAF index for the superior WGEP model was increased almost three times after using the mother wavelet. Furthermore, the R and MARE statistical indices for the WGEP model were computed to be 0.935 and 0.862, respectively.

Rainfall-runoff process is one of the most important and complex phenomena in the hydrological cycle. Therefore, in its modeling, different perspectives for the development and improvement of predictive models have been presented. In this study, while introducing a combination of wavelet-group classification of data, its effectiveness for modeling the rainfall-run-off process in the Ghara-eos watershed was studied. At first, the rainfall and runoff time series were decomposed using a wavelet transform to several sub-basins to overcome its non-state. Then, these time subcircuits are considered as inputs of the grouped data collection method for predicting daily runoff. The efficiency of the combined model with DC and root mean square error (RMSE) were evaluated. The results of the validation of the models indicate that the highest amount of explanation coefficient and the lowest root mean of error for the single GMDH model were 0.65 and 0.07, respectively, and for the combined model of grouping the wavelet data The order is 0.91 and 0.05. The reason for the hybrid model's superiority to the single model is that the combination model of the grouping of wavelet data, instead of using the time series of rainfall and runoff data on a general scale, uses several time- Different decompositions are used as inputs in the model. Also, the results showed that the Wavelet-GMDH combination model compared to other composite models such as Waveline Artificial Neural Network (WANN) due to the GMDH model layer function, which includes binary combinations of input variables, and by selecting the number of optimal neurons in Each layer directs motion to the predicted data, has more efficiency and accuracy.

In present study, the groundwater level of the Kabodarahang region located in Hamadan Province was simulated using novel techniques such as Self-Adaptive Extreme Learning Machine (SAELM) and Wavelet-Self-Adaptive Extreme Learning Machine (WA-SAELM). Firstly, the effective lags were detected using the autocorrelation function and then ten models were developed for each SAELM and WA-SAELM methods. By evaluating the results of the models, WA-SAELM was introduced as the superior method. The analysis of the simulation results showed that the superior model had a high accuracy in estimating the groundwater level. For the superior model, the correlation coefficient (R), Root Mean Squared Error (RMSE) and Nash-Sutcliffe efficiency coefficient (NSC) were calculated to be 0.969, 0.358 and 0.939, respectively.

In this study, the groundwater level (GWL) of the Sarab Qanbar region located in the south of Kermanshah, Iran, was estimated using the Wavelet- Self- Adaptive Extreme Learning Machine (WA- SAELM) model. An artificial intelligence method called “Self- Adaptive Extreme Learning Machine” and the “Wavelet transform” method were implemented for developing the numerical model. First, by using the autocorrelation function (ACF), the partial autocorrelation function (PACF) and the effective lags in estimating GWL, eight distinctive SAELM and WA- SAELM models were developed. Later, the values of the observational well were normalized for estimating GWL. Next, the most optimized mother wavelet was chosen for the modeling. By evaluating the results of SAELM and WA- SAELM, it was concluded that the WA- SAELM models could estimate the values of the objective function with higher accuracy. Then, the superior model was introduced, showing that it could be very accurate in forecasting the GWL. In the test mode, for example, the values of R (correlation coefficient), Main absolute error (MAE) and the NSC- Sutcliffe efficiency coefficient (NSC) for the superior model were calculated to be 0.995, 0.988 and 0.990, respectively. Furthermore, an uncertainty analysis was conducted for the numerical models, proving that the superior model had an underestimated performance.

The need to simulate rainfall time series at different scales for engineering purposes on the one hand and lack of recording such parameters in small scales because of administrative and economic problems, on the other hand, disaggregation of rainfall time series to the desired scale is an essential topic. In this study, for disaggregating the Tabriz and Sahand rain gauges time series, according to nonlinear characteristics of time scales, wavelet- Least Square Support Vector Machine (WLSSVM) hybrid model is proposed and daily data of four rain gauges and monthly data of six rain gauges from Urmia Lake Basin for ten years were decomposed with wavelet transform and then by using mutual information and correlation coefficient criteria, the subseries were ranked and superior subseries were used as input data of Least Square Support Vector Machine (LSSVM) model for disaggregating the Tabriz and Sahand rain gauges monthly rainfall time series to the daily time series. Results obtained from the WLSSVM disaggregation model were compared with the results of LSSVM and traditional multiple linear regression models. The results of WLSSVM model to LSSVM and multiple linear regression models at validation stage in the optimized case for Tabriz rain gauge were increased 10% and 37.5% and in the optimized case for Sahand rain gauge were increased 24.5% and 46.7% respectively. It was concluded that hybrid WLSSVM model has a higher accuracy than two other methods and can be considered as an accurate disaggregation model to disaggregate the rainfall time series.

Without a doubt, the first step to manage water resources is accurately predicting of river flow. In This study, to predict time series of daily and monthly of kakareza and sarab seyed ali stations, artificial neural network and Adaptive neuro-fuzzy inference system models is used. the daily and Monthly rainfall and runoff time series of Sarab seyedali and kakareza basins were decomposed into sub-signals in various resolution levels using wavelet analysis and then these sub-signals entered to the ANN and ANFIS models to reconstruct the main forecasted time series. The model results show the high merit of Db-4 wavelet in comparison with the Haar wavelet. Too, results showed that WANFIS model produced better forecasts than either the ANN, WANN, ANFIS models.